Understanding Recovery and Backup on Cloud - Moving to a cost effective convenient solution

Article by Karan Kirpalani



If there is one thing the public cloud has proven adept, it is providing a low-cost means to house all the data that both large and small enterprises generate on a daily basis.

But even more importantly, it provides a convenient means to ensure data access and business continuity in the event of a serious outage. Now, instead of mirroring a complete data environment in a remote datacenter somewhere, firms can lease the necessary resources at low cost, and then shift loads over to the cloud should primary systems fail.

Keep in mind, however, that the cloud represents a substantial shift in the nature of IT infrastructure, leading to fundamental issues regarding data ownership, service responsibilities and infrastructure management. That means the cloud is not only a technical challenge, but an operational one that requires some hard decisions to be made as to how it is to be provisioned, designed and put into service.

One of the more crucial issues is the type of cloud architecture required to suit your needs. The three most popular are Software-as-a-Service (SaaS), Platform-as-a-Service (PaaS) and Infrastructure-as-a-Service (IaaS), each of which provide a different level of reliance on both internal and external infrastructure. SaaS requires the least dependency on outside resources, as it is designed to deliver application level services like Business Intelligence (BI) and Customer Resource Management (CRM). PaaS is primarily an application development and deployment environment, while IaaS provides end-to-end infrastructure support and as such is most suited to complete enterprise restoration in the event of a significant outage, say, from a natural disaster.

Of course, backup and recovery solutions are of no use if they fail to activate during an emergency. That's why it is important to both define and then verify reliability claims before a crisis emerges. Evaluating reliability should encompass four elements: hardware/software infrastructure, cloud provider personnel, service connectivity and enterprise personnel. Each of these components needs to be prepped and ready to shift to an all-cloud basis should the worst occur, which, again, means both infrastructure and procedures need to be clearly defined and properly tested ahead of time.

The cloud introduces a number of thorny issues surrounding data ownership as well. Once data has left the confines of the enterprise infrastructure, a host of legal issues arise depending on where, exactly, data resides. Requirements for data discovery, compliance, visibility and numerous other functions vary from country to country, and quite often within multiple regions of a single company. And in many cases, legal standards are still evolving as court systems and legislators around the world attempt to keep pace with the rapid advancement of technology. The best protection, then, is a clear-cut, written agreement between enterprise and cloud provider, detailing exactly where and how backup infrastructure will be maintained and then put into action. And remember, many cloud providers have their own backup and recovery facilities, most likely using additional cloud services, so it is crucial for enterprises to know where their data resides at all times.

It is also important to remember that data outages are not necessarily the result of major natural disasters like hurricanes and typhoons. The most common cause, in fact, is human error, followed by system failure. And outages are not necessarily total, with partial service disruption or high application latency also affecting data operations and, ultimately, enterprise revenues.

The good news is the cloud provides an effective, reasonably priced backup and recovery solution, enabling even small organizations to gain peace-of-mind regarding their data infrastructure. But with the cloud evolving at a rapid pace, it is important for enterprises to realize that backup and recovery in the cloud is an ongoing process involving system tests, policy updates, performance evaluations and connectivity/data migration assessments.

Enterprises that fail to ensure their recovery and business continuity measures are in good working order will discover the inadequacies only in the event of an actual emergency − when it is too late to do anything about it.